Skip to content

Fable5: Transcript .txt file omits speaker names #18

@etiennechabert

Description

@etiennechabert

Problem

The live UI shows speaker names, but the saved transcript .txt file contains only timestamps and text — speaker names captured via the Meet bot (or diarization labels) never make it into the file:

[2026-06-10 14:30:12] [0.00s-4.20s] Hello everyone, welcome to the meeting

Root cause

In app.pytranscribe_and_translate(), the per-segment speaker label is computed (seg['speaker'], with the real name when caption resolution succeeded) and even printed to the console — but the f.write() call for the transcript file omits it.

Fix

Include the speaker label in each written line:

[2026-06-10 14:30:12] [0.00s-4.20s] Alice Smith: Hello everyone, welcome to the meeting

The summarization parsers (load_transcript_segments, run_pending_summary) use a (.+) text group, so both old- and new-format files still parse.

Status

Fixed in PR #16.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions